A Two-Tube Multiplex Reverse Transcription PCR Assay for Simultaneous Detection of Sixteen Human Respiratory Virus Types/Subtypes

There is a need for the development of a rapid and sensitive diagnosis of respiratory viral pathogens. With an intended application in provincial Centers for Diseases Control and Prevention, in this study, we present a two-tube multiplex RT-PCR assay (two-tube assay) using automatic electrophoresis to simultaneously detect sixteen common respiratory viruses. The specificity and the sensitivity of the assay were tested. The assay could detect 20–200 copies per reaction when each viral type was assayed individually, 2000 copies with 9 premixed viral targets in the multiplexed assay in tube 1, and 200 copies with 8 premixed templates in tube 2. A total of 247 specimens were used to evaluate the two-tube assay, and the results were compared with those obtained from the Luminex xTAG RVP Fast assay. The discordant results were confirmed by sequencing or by the Seeplex RV15 ACE detection kit. There were no false positives, but six false negatives occurred with the two-tube assay. In conclusion, the two-tube assay is demonstrated to have great potential for routine surveillance of respiratory virus infection in China.


Introduction
Viral respiratory tract infections are a common cause of hospitalization and mortality in infants and young children. Influenza A, influenza B, parainfluenza, human rhinovirus, adenovirus, respiratory syncytial virus, metapneumovirus, and coronavirus are recognized as viral agents associated with respiratory tract infections [1][2][3]. Therefore, there is a need for a rapid and sensitive differential diagnosis of respiratory viral pathogens to prevent nosocomial infections and to minimize unnecessary antibiotic usage [4][5][6].
Existing standard nonmolecular diagnostic methods such as viral culture and immunofluorescence (DFA) are timeconsuming, labor intensive, or have low sensitivity [7,8]. Accordingly, molecular techniques are increasingly accepted for diagnosis of viral respiratory tract infections. In the last few years, multiplex RT-PCR assays have been developed to detect respiratory viruses [9], and some have been commercialized, such as xTAG RVP from Luminex [10,11], Multicode PLx RVP from EraGen Biosciences [12], and ResPlex II from Qiagen [13].
In our previous study, a multiplex PCR based on the GenomeLab Gene Expression Profiler (GeXP) analyzer had been shown to be effective for the detection of pandemic influenza A H1N1 virus [14], nine serotypes of enteroviruses associated with hand, foot, and mouth disease [15], and sixteen different respiratory virus types/subtypes in a single tube [16]. However, these methods are not likely to be widely adopted in provincial Centers for Disease Control and Prevention due to the limited availability of GeXP equipment. Therefore, in this study, a two-tube multiplex reverse transcription PCR assay (twotube assay) to detect sixteen respiratory viruses based on the amplicon size differences using automated electrophoresis system is described. One tube is used for the simultaneous detection of nine respiratory viruses, including influenza A virus (FluA), influenza B virus (FluB), seasonal influenza A virus subtypes H1N1 (sH1N1), parainfluenza virus type 1 (PIV1), human rhinovirus (HRV), coronavirus subtypes OC43 (CoV OC43), coronavirus subtypes 229E (CoV 229E), coronavirus subtypes HKU1 (CoV HKU1), and adenovirus (Adv). Another tube is used for the simultaneous detection of seven respiratory viruses including parainfluenza virus type 2 (PIV2), parainfluenza virus type 3 (PIV3), respiratory syncytial virus A (RSVA), respiratory syncytial virus B (RSVB), coronavirus subtypes NL63 (CoV NL63), human metapneumovirus (HMPV), and human bocavirus (HBoV). As the two-tube assay uses the QIAxcel automated electrophoresis system, which is accessible in most of provincial Centers for Disease Control and Prevention in China, this two-tube assay may have greater potential for the routine surveillance of respiratory virus infection in China and to improve the capacity for emergency management.

Specimens.
A total of 247 nasopharyngeal aspirates were collected in the Children's Hospital of Beijing, China, during June 2008 and March 2010 from hospitalized children under two years of age presenting acute lower respiratory infection (ALRI) syndromes based on clinical measurements recommended by the World Health Organization (WHO). The patients were included in this study according to the following criteria: parenchymal infiltration on chest radiography, dry or moist rale, body temperature above 37.5 ∘ C, a normal or low leukocyte count, or lower chest wall indrawing. The mean age of study participants was 13.08 months. Of 247 cases, 149 (60.32%) participants were male and 98 (39.68%) female. A total of 160 (64.78%) participants were from urban areas and 87 (35.22%) were from suburban or rural areas of China (Hebei, Henan, Shanxi, Shandong, and Inner Mongolia). The specimens were collected within 5 to 7 days after symptoms onset. Patients with bacterial infections were excluded. All aspects of the study were performed in accordance with national ethics regulations and approved by the Institutional Review Boards of the Center for Disease Control and Prevention of China. Children's parents were apprised of the study's purpose and of their right to keep information confidential. Written consent was obtained from children's parents.
Total RNA/DNA was extracted from 200 L of clinical sample using the QIAamp Viral RNA Mini Kit (Qiagen, Hilden, Germany). The extracts were eluted into 50 L of DNase-and RNase-free water and stored at −80 ∘ C.

Primers.
Nine pairs of chimeric primers [17] were added to tube 1 to detect nine respiratory viruses, and eight pairs of chimeric primers were added to tube 2 to detect seven respiratory viruses. One pair of internal control primer and one pair of universal primers were added to both tubes. The sequence from human genome RNase P gene was used as an internal control primer for the specimens [16]. The primers sequences, the target genes, the amplicon sizes, and primer working concentrations are listed in Table 1.

Two-Tube Assay and Detection. One-
Step RT-PCR Kit (Qiagen, Hilden, Germany) was used for the amplification. A total of 25 L PCR mixture containing 2 L of extracted RNA and varied primer concentrations (Table 1) was subjected to the following conditions: 50 ∘ C for 30 min, 95 ∘ C for 15 min, followed by 10 cycles of 95 ∘ C for 30 s, 55 ∘ C for 30 s, and 72 ∘ C 30 s; 10 cycles of 95 ∘ C for 30 s, 65 ∘ C for 30 s, 72 ∘ C for 30 s; 25 cycles of 95 ∘ C for 30 s, 48 ∘ C for 30 s, and 72 ∘ C for 30 s, and a final incubation of 72 ∘ C for 3 min [15], and the reactions were analyzed on the QIAxcel automatic electrophoresis using QIAxcel DNA High-Resolution kit.

Specificity and Sensitivity of the Two-Tube Assay.
The specificity of the two-tube assay on all viral targets was tested individually in the two multiplex reactions under the experimental condition as described previously. Cell-cultured virus stocks contained in NATtrol Respiratory Validation Panels 2 (NATRVP-2) (ZeptoMetrix, Buffalo, New York, USA) were used as the positive controls of FluA, FluB, sH1N1, PIV1-3, HRV, Adv, RSVA, RSVB, HMPV, CoV 229E, and CoV OC43 to validate the specificity. Specimens genotyped previously [18,19] were used as positive controls of HBoV, HKU1, and CoV NL63.
To determine the analytical sensitivity of the assay, serial dilutions of viral RNA transcripts of all RNA viruses and serial dilutions of the plasmids containing the Adv and HBoV target sequence were analyzed by the two-tube assay. The PCR products were cloned into a pGEM-T vector, which was used to transform DH10B cells. The plasmid DNA was extracted with an E.Z.N.A. Plasmid Mini Kit I (Omega, GA, USA). The RNA copy number was calculated after measuring the concentration of the RNA transcribed in vitro using a T7 Large-Scale RNA Production System (Promega, Wisconsin, USA). For the DNA viruses Adv and HBoV, the in vitro transcription was omitted. The sensitivity of the two-tube assay was examined by using premixed quantitative viral RNA transcripts or plasmids. The personnel responsible for reading these and other assays were experienced staff with basic training in microbiology and molecular biology.

Resolution and Confirmation of Discordant Results.
The additional targets detected only by the two-tube assay were confirmed by independent PCR and sequencing.

Specificity and Sensitivity of the Two-Tube Assay.
The expected size of each virus type/subtype-specific amplicon was observed and separated clearly from the other viral targets on QIAxcel automatic electrophoresis for all of the positive controls. No mispriming (primer dimer) or other amplification in the negative controls was observed in either tube ( Figure 1). The QIAxcel DNA High Resolution kit is capable of resolving amplicons with as little as 5 bp size difference; the smallest size difference in this test was 10 bp.
For the viral targets that used clinical specimens as positive controls (HBoV, CoV HKU1, and CoV NL63), a single PCR product was detected in addition to the internal control peak (126 bp). The sensitivity of the assay was evaluated for each virus type/subtype individually using serial tenfold dilutions ranging from 10 to 10 5 copies of cloned PCR products in both of the two tubes. The limit of detection for HRV, PIV2, PIV3, RSVA, HBoV, and Adv was 20 copies per reaction and 200 copies per reaction for the other 10 virus type/subtypes assayed in this study.
The sensitivity of two-tube assay when all the viral targets were present was also tested. The detection sensitivity in tube 1 with 9 pre-mixed viral targets was 2000 copies per reaction and 200 copies per reaction in tube 2 with 7 premixed templates ( Figure 2).

Evaluation of the Two-Tube Assay Using Respiratory
Specimens. All the comparative detections were double-blind tests performed by trained staff in our laboratory. All of the 247 specimens assayed with the Luminex xTAG RVP Fast assay were retested by the two-tube assay. The two assays detected presence of at least one of the 16 assayed viruses in 194 specimens, and the two assays both detected virus in 181 specimens. The two-tube assay detected virus in 8 specimens that were negative in the Luminex xTAG RVP Fast assay (2 Adv, 3 RSVB, 2 HRV, and 1 PIV3), and 5 specimens had a positive detection of HRV only by the Luminex xTAG RVP Fast assay. The Luminex xTAG RVP Fast assay identified 111 specimens to be coinfections, 83 of them (74.77%) were in complete agreement between both assays, and detection of 13 viruses (in 12 specimens) was missed by the two-tube assay. Additional 14 viruses (in 14 coinfected specimens) were detected only by the two-tube assay, and 2 coinfection specimens (no. 28 and no. 118) gave inconsistent results between the two assays ( Table 3). As shown in Table 2, there were 40 viral infections identified only by the two-tube assay and 21 viral infections found only by the Luminex xTAG RVP Fast assay. The infection events not detected by the Luminex xTAG RVP Fast assay included PIV3 ( = 2), HRV ( = 15), HMPV ( = 1), Adv ( = 11), RSVA ( = 1), and RSVB ( = 10). Thirty-two of these undetected targets were found in coinfections, and the remaining 8 targets (2 Adv, 3 RSVB, 2 HRV, and 1 PIV3) were single-target positive specimens by the two-tube assay. The two-tube assay failed to detect the following viral infections ( = 21): sH1N1 ( = 2), HRV ( = 13), Adv ( = 1), and HBoV ( = 5). Sixteen of the 21 specimens were coinfections, and the remaining 5 targets were positive only for HRV specimens. Because the HRV primers used in the Luminex xTAG RVP Fast assay were able to amplify both HRV and enterovirus, the specimens positive for HRV detected only by the Luminex xTAG RVP Fast assay could be enteroviruses. The sensitivity, specificity, negative prediction value (NPV), positive prediction value (PPV), the accordance rate, and the kappa value of each virus for the twotube assay, when compared to the Luminex xTAG RVP Fast assay as a reference, are shown in Table 2.

Resolution and Confirmation of Discordant Results.
All of the 40 additional targets detected only by the two-tube assay were confirmed by sequencing as true positives. For those specimens with discordant results by the two-tube and RVP assays, the Seeplex RV15 ACE detection kit was used to resolve the discrepancy ( Table 3).
The confirmed results were defined as the results detected by at least two of the three PCR assays [22,23], or the results of sequencing. The sensitivity, specificity, negative prediction value (NPV), positive prediction value (PPV), the accordance rate, and the kappa value of each virus for the two-tube assay, when compared to the confirmed results as a reference, are shown in Table 4.

Discussion
In this study, a novel two-tube assay based on the QIAxcel capillary electrophoresis to replace GeXP has been evaluated for the potential application in routine laboratory testing of respiratory viruses in China. Under optimized working condition, the two-tube assay revealed good specificity without nonspecific amplification using size detection system though conventional validation of specificity is usually done by probe detection, hybridization, nested PCR, and sequencing. The two-tube assay achieved a sensitivity of 20-200 copies per reaction when assayed individually for each virus type/subtype, and 2000 copies when 9 viral targets were present in tube 1 and 200 copies when 7 viral templates were  present in tube 2. These results suggest the utility of the two-tube assay for the detection of multiple respiratory virus infections for laboratories not equipped to perform liquid chip-based multiplex assays. The 247 specimens from hospitalized children showed a high prevalence of infection and co-infection with the common respiratory viral pathogens. HRV was found most frequently, followed by RSVB. The major discrepancies between the two-tube assay introduced in this study and the Luminex xTAG RVP Fast assay were found for the detection of HRV, Adv, HBoV, RSVB, and sH1N1 (kappa = 0.770, 0.773, 0.907, 0.913, and 0.929, resp.) ( Table 2). All of the HRV positive samples including 15 detected only by twotube assay were confirmed by sequencing to be positive for HRV A. The 13 HRV-positive specimens detected only by the Luminex xTAG RVP Fast assay could be enteroviruses because of the Luminex RVP Fast assay detecting enterovirus (the HEV component of the HEV/HRV). After resolved by the Seeplex RV15 ACE detection kit, three of them were truly discordant. For Adv, seven specimens identified as positive only by the two-tube assay were shown via sequencing to be Adv C, and 4 were Adv B1. Only one Adv-positive specimen was missed by two-tube assay. For HBoV, five specimens detected as positive only by the Luminex xTAG RVP Fast assay were also negative by the Seeplex assay. The other 14 positive samples detected only by two-tube assay (10 RSVB, 2 PIV3, 1 RSVA, and 1 HMPV) were confirmed by sequencing as true positives. For sH1N1, all of the FluA positives were sequenced, and 16 of 20 were sH1N1, the two-tube assay failed to detect 2 of the 16 sH1N1 positives. The detection results of coronaviruses (CoV HKU1, CoV NL63, CoV 229E, and CoV OC43), influenza (FluA, FluB), PIV1, and PIV2 were completely consistent between the two assays. In summary, no false positives were found by the two-tube assay, and a total of 6 false negatives occurred: HRV ( = 3), sH1N1 ( = 2), and Adv ( = 1). The two-tube assay was more sensitive than the Luminex xTAG RVP Fast assay for the detection of HRV A, RSVB, Adv B1, and Adv C. The overall detection rate of the two-tube assay for each virus was comparable to that of the Luminex xTAG RVP Fast assay (kappa > 0.75) demonstrating the high sensitivity and specificity of the twotube assay in the analysis of clinical samples.
The two-tube assay, as confirmed by the Luminex xTAG RVP Fast assay, revealed that almost 80% of the respiratory specimens tested were positive for at least one viral infection, and viral coinfections were frequent. The overall co-infection rates of the two-tube assay and the Luminex xTAG RVP Fast assay were 48.58% (120/247) and 44.94% (111/247), respectively. Compared with other similar reports [9,[24][25][26], the extremely higher co-infection rate found by both assays might be due to the difference in the tested population and possible nosocomial infections in the hospital.
The routine use of extensive testing by methods such as the two-tube assay should make large studies of the clinical relevance of single and multiple respiratory virus infections more feasible. The clinical relevance of the outcome of multiplex PCR tests for respiratory viruses is not yet fully determined. In one study of children with RSV infection, higher fever, longer hospital stays, and more frequent use of antibiotics were associated with multiple infections [27]. For some agents, such as influenza A and B virus, a positive test may provide the basis for antiviral treatment [28]. It has been suggested that a rapid etiologic diagnosis of viral RTI could reduce unnecessary prescription of antibiotics, but this remains to be shown [28].
Two distinct advantages of the two-tube assay are the short assay time and the low cost. The assay requires 5-6 hours to complete 100 samples, including a half hour to prepare PCR mixture, the whole RT-PCR (3 hours), and the detection on the QIAxcel automatic electrophoresis (12 tests/15 minutes), while using other liquid chip-based assays might take an entire workday [29]. The cost for two-tube assay is $10/test, including the RT-PCR kit and the consumables of detection versus at least $120/test [29] using other liquid chip-based assays [29].
In conclusion, the two-tube assay developed in this study using automatic capillary electrophoresis is a rapid, cost-effective method with high sensitivity and specificity for the detection of respiratory virus infection, and it is demonstrated to have great potential for routine surveillance of respiratory virus infection.